Slowing and Storing Microwaves in a Single Superconducting Fluxonium Artificial Atom

Three-level Lambda systems provide a versatile platform for quantum optical phenomena such as Electromagnetically Induced Transparency (EIT), slow light, and quantum memory. Such Lambda systems have been realized in several quantum hardware platforms including atomic systems, superconducting artificial atoms, and meta-structures. Previous experiments involving superconducting artificial atoms incorporated coupling to additional degrees of freedom, such as resonators or other superconducting atoms. In this work, we performed an EIT experiment in microwave frequency range utilizing a single Fluxonium qubit within a microwave waveguide. The Lambda system is consisted of two plasmon transitions in combination with one metastable state originating from the fluxon transition. In this configuration, the controlling and probing transitions are strongly coupled to the transmission line, safeguarding the transition between 0 and 1 states, and ensuring the Fluxonium qubit is close to the sweet spot. Our observations include the manifestation of EIT, a slowdown of light with a delay time of 217 ns, and photon storage. These results highlight the potential as a phase shifter or quantum memory for quantum communication in superconducting circuits.

💡 Research Summary

In this work the authors demonstrate electromagnetically induced transparency (EIT), slow‑light propagation, and photon storage using a single superconducting fluxonium artificial atom directly coupled to a three‑dimensional copper waveguide. By biasing the fluxonium near its sweet spot (φ_ext/φ₀≈0.53) they engineer a Λ‑type three‑level system in which the two “plasmon” transitions (|0⟩↔|2⟩ and |1⟩↔|2⟩) lie within the waveguide passband, while the metastable “fluxon” transition (|0⟩↔|1⟩) is placed in the stop band with >40 dB isolation. This configuration allows both the control field (driving |1⟩↔|2⟩) and the weak probe field (driving |0⟩↔|2⟩) to be strongly coupled to the transmission line, yet protects the low‑frequency transition from radiative loss, preserving coherence.

The dynamics are modeled with a Born‑Markov master equation that includes relaxation rates Γ_ij and pure dephasing γ_jj for all three levels. The effective Hamiltonian in the rotating frame contains the probe and control Rabi frequencies (Ω_p, Ω_c) and detunings Δ_p, Δ_c. The steady‑state transmission coefficient t is obtained via the input‑output relation t = 1 + iΓ_02 Ω_p ρ_02. Numerical solutions using QuTiP reproduce the measured transmission spectra with high fidelity.

Two‑tone spectroscopy reveals a clear transition from the EIT regime to an Autler‑Townes splitting (ATS) regime as the control power increases. The boundary is quantified by Ω_EIT = γ_02 – γ_01; for Ω_c < Ω_EIT the system exhibits a narrow transparency window, whereas Ω_c > Ω_EIT produces a doublet characteristic of ATS. The authors fit the control‑field power dependence of the effective Rabi frequency and confirm the expected √P scaling.

Phase measurements of the transmitted probe allow extraction of the group delay τ_d = –∂Arg(t)/∂ω_p. At a control Rabi frequency of Ω_c/2π = 2.6 MHz and zero probe detuning, a maximum delay of 217 ns is observed, corresponding to a reduction of the group velocity by several orders of magnitude. By slightly detuning the probe, the phase slope reverses, giving rise to a fast‑light (superluminal) effect, which the authors also document.

For photon storage, a short Gaussian probe pulse (σ = 0.05 µs) with an average photon number ≈0.006 is sent into the waveguide. While the control field is on, the pulse experiences EIT‑induced delay; the control is then switched off, mapping the photonic excitation onto the long‑lived |1⟩ state of the fluxonium. After a storage interval τ_s = 0.5 µs the control field is turned back on, retrieving the pulse. The measured storage‑retrieval efficiency reaches 12 %, limited primarily by the dephasing rate γ_01 of the fluxon transition and the fact that the Λ system is operated slightly away from the optimal sweet spot. The authors estimate that, if the system were tuned exactly at the sweet spot, the delay could be increased to ~604 ns.

The paper provides a comprehensive characterization of the fluxonium parameters (E_J = 9.041 GHz, E_C = 0.995 GHz, E_L = 0.807 GHz) extracted from one‑tone spectroscopy and fits to the three‑level model. The device fabrication details (Al/AlOx/Al junction, superinductance array of 180 large junctions, bow‑tie antenna) and the waveguide geometry (TE₁₀ mode, electric field polarization orthogonal to the dipole direction) are described, highlighting the careful engineering that enables strong atom‑waveguide coupling while suppressing unwanted loss channels.

In summary, the authors achieve EIT, slow‑light, fast‑light, and photon‑storage functionalities with a single fluxonium qubit, without auxiliary resonators or additional qubits. This minimalist architecture simplifies fabrication, reduces decoherence pathways, and offers a scalable platform for microwave quantum optics, quantum networking, and on‑chip quantum memory applications in superconducting circuits.